The four tiers
A category's published headline uses the highest tier that has cleared the input-reality gate (Pass 0). Tier-A always wins when it exists. Tier-C is the foundation that survives until Tier-A clears.
Tier A — anchor
Verified transactions
Stripe-confirmed transactions on the WorkForce marketplace. Real cost, real eval execution, real measured buyer outcomes. The only tier that defines a true WLI rate.
currently: 0 categories · pre-TX1
Tier B — feed
Signed third-party feeds
Buyer-attested pricing under a signed data-share agreement (procurement teams, marketplaces, industry consortia). Higher quality than list price but not transaction-anchored.
currently: 0 categories · agreements in progress
Tier C — reference
Public vendor list prices
Simple median of public vendor pricing pages, range across N≥4 publishable vendors per category. Custom / contact-sales vendors are excluded (opacity is recorded, never estimated).
currently: 10 categories · live below
Tier D — opinion
Analyst estimates
Forrester / Gartner / individual-blog estimates without a vendor-attributable source. Excluded by default. Never used in a published WLI figure.
excluded · methodology-only reference
What Tier-C is and is not
Tier-C is the simple median of a category's public vendor list prices, with the min-max range of the per-vendor inputs and the count of contributing vendors. It is a robust, reproducible reference — anyone can re-derive it from the same source URLs. The figure is honestly labeled Tier-C · N vendors in the /wli table. It is not a confidence interval (CIs require transaction-anchored data — Tier-A). It is not what buyers actually pay after negotiation — negotiated procurement prices typically discount the published list by 20–50%, and that discount only becomes visible at Tier-A.
Auto-promotion: Tier-C → Tier-A
A category promotes from Tier-C to Tier-A when all three fire: (1) n≥50 verified Stripe-confirmed transactions land in the marketplace database for that category, (2) each transaction has a real eval_score (from a non-simulated eval engine) and a real outcome_score (buyer attestation or marketplace-derived success signal), and (3) the 3-pass pressure test re-passes on the resulting median. On promotion, the Tier-C number is preserved as historical reference on the per-category page; the headline switches to Tier-A. No auto-demotion: a Tier-A rate freezes (with a "stale" marker) rather than silently falling back to Tier-C — protecting the citation contract.
Per-category Tier-C source tables
Every Tier-C figure currently rendered on /wli, with the full per-vendor source data. Anyone can re-derive the median from these URLs.
CS Ticket Resolution
$0.50 · [0.10–2.00]
Unit note. Outcome-class unit covers per-resolution, per-resolved-conversation, per-session, per-conversation, per-outcome. Per-outcome (Fin), per-resolved-conversation (HubSpot), per-automated-resolution (Zendesk) fire on success; per-session (Freshworks) and per-conversation (Salesforce) fire on usage. NOT a like-for-like unit; median reflects the mixed outcome-class denomination as actually charged.
Excluded (opaque / contact-sales): Sierra; Decagon; Ada; Forethought; Aisera; Cresta; Parloa; Cognigy; Kore.ai; Zendesk ($-not-published).
Source file: research/output/cs-resolution-published-prices_2026-06-19.md
Code Generation
$22.50 · [10.00–39.00]
Unit note. Per-seat / month is the dominant unit across published AI-code-generation tools. Lowest paid self-serve tier used per vendor; annual-discounted unit price where shown. Per-seat is NOT a per-task price — Tier-A will eventually move to per-task once transaction data lands.
Excluded (opaque / contact-sales): Sourcegraph Cody (enterprise-only); Lovable (official $ not machine-readable).
Source file: research/output/code-generation-published-prices_2026-06-19.md
Code Review (PR)
$27.00 · [15.00–39.00]
Unit note. Per-developer / month is the published unit across self-serve AI code-review tools. Per-PR pricing exists only on enterprise tiers behind a sales gate; the Tier-C median reflects the self-serve per-seat clearing rate.
Excluded (opaque / contact-sales): Snyk Code (enterprise quote); Semgrep (enterprise quote); Sentry Bits (bundled); Datadog Bits (custom); DeepSource (enterprise gated); CodeGuru (AWS usage-based, opaque).
Source file: research/output/tier-c-medians_2026-06-19.json
Call Transcription
$14.00 · [7.00–20.00]
Unit note. Per-user / month entry paid tier across the AI-notetaker cohort. Annual-billed unit prices used where shown.
Excluded (opaque / contact-sales): tldv (JS-rendered, not verbatim-fetched); Zoom AI Companion (bundled with paid Zoom).
Source file: research/output/meeting-assistant-published-prices_2026-06-19.md
Meeting Summary
$14.00 · [7.00–20.00]
Unit note. Same vendor cohort as call-transcription. Meeting-summary is the higher-intent sales/RevOps buyer query; underlying SaaS is identical.
Excluded (opaque / contact-sales): tldv; Zoom AI Companion.
Source file: research/output/meeting-assistant-published-prices_2026-06-19.md
Document Translation
$17.50 · [10.00–25.00]
Unit note. Cloud MT API rate normalised to per-1M-characters. Two sub-units coexist in the broader cohort (MT API per character; localization-platform per seat/month) — the Tier-C published here covers ONLY the MT API rate. Per-seat localization platforms (Phrase, Weglot, Lokalise) are a different product and excluded.
Excluded (opaque / contact-sales): Lilt (enterprise custom); Unbabel (enterprise custom); Smartling (managed-translation, different unit); Phrase / Weglot / Lokalise (localization platform, different product).
Source file: research/output/translation-published-prices_2026-06-19.md
Short-form Content
$34.00 · [7.50–59.00]
Unit note. Self-serve entry paid tier per seat/month. Annual-billed unit prices where shown. Bundled-seat enterprise tiers (Copy.ai Growth at $1000/mo across 75 seats) excluded — per-seat economics not directly comparable.
Excluded (opaque / contact-sales): Hypotenuse AI (no $ on official page); Copy.ai (bundled-seat tiers, not directly comparable).
Source file: research/output/marketing-content-published-prices_2026-06-19.md
Knowledge-base Query
$10.00 · [7.25–30.00]
Unit note. Per-user / month for AI-knowledge-search bundled tier or add-on. Annual-billed prices used where shown.
Excluded (opaque / contact-sales): Glean (contact-sales); Moveworks (custom quote); Guru (quote-based); Coveo (entitlement / quote).
Source file: research/output/knowledge-search-published-prices_2026-06-19.md
Lead Qualification
$134.00 · [49.00–250.00]
Unit note. MIXED UNIT cohort: Apollo per-seat; Clay per-workspace with credit quota; AiSDR / Artisan / Lindy per-AI-worker subscription. Tier-C reflects the entry paid headline rate per vendor — NOT a per-lead price.
Excluded (opaque / contact-sales): 11x Alice (enterprise-scale outlier); Outreach (enterprise quote); Salesloft (enterprise quote).
Source file: research/output/sdr-lead-qualification-published-prices_2026-06-19.md
Email Drafting
$25.00 · [9.00–30.00]
Unit note. Per-user / month entry paid tier for AI-augmented email composition. Vendors share a common per-seat unit, making the median honest.
Excluded (opaque / contact-sales): Smartwriter (per-credit, different unit); Compose AI (free-tier focus).
Source file: vendor pricing pages (consolidated into tier-c-medians_2026-06-19.json)
Citing a Tier-C rate
Recommended format: Workforce Labor Index, [category], Tier-C reference rate v1.0, [date], n=[N] vendors. Always include the tier and the vendor count — a Tier-C rate cited without "Tier-C" risks being read as transaction-anchored, which it is not.
Pushback channel: index@workforcebygriffain.com. Wrong source URL, wrong price, missing vendor, opaque-but-now-published — write in and the row updates with a changelog entry.
← back to /wli · IOSCO statement of compliance →